Retrieval of boreal forest parameters from CARABAS image texture

The intensity of CARABAS SAR imagery has previously been used for stem volume retrieval in oak dominated forests. Lately the stem volume of boreal forests has been examined using the backscattered intensity as the starting point. The work presented here is based on studying the standard deviation of the backscattering coefficient /spl sigma//sup 0/ rather than the intensity itself. The logarithm of the standard deviation of /spl sigma//sup 0/ is regressed to the logarithm of the distance within which the standard deviation is calculated. In practice this is done using a block of 11/spl times/11 pixels so that the standard deviation of /spl sigma//sup 0/ is averaged for all horizontal and vertical alternatives of equal length. The constant term and the coefficient of the linear term of the regression are the tools used for forest parameter retrieval.